Jürgen Berges Darmstadt University of Technology

19
QCD Plasma Instabilities and Nonthermal Fixed Points Jürgen Berges Darmstadt University of Technology

Transcript of Jürgen Berges Darmstadt University of Technology

QCD Plasma Instabilities

and

Nonthermal Fixed Points

Jürgen Berges

Darmstadt University of Technology

Content

• Nonequilibrium instabilities (Weibel, Nielson-Olesen)

• Classical-statistical lattice gauge theory

• Characteristic time scales of SU(2), SU(3) pure gauge

• Instability-induced fermion production

• Kolmogorov wave turbulence vs. infrared fixed points

Nonequilibrium dynamics

Relativistic heavy-ion collisions explore

strong interaction matter starting from a

transient nonequilibrium state

Thermalization process?

Schematically:

• Characteristic nonequilibrium time scales? Relaxation? Instabilities?

tiso » O(1/gT)

Isotropization time tiso? In the absence of nonequilibrium instabilities:

tiso » O(1/g4T)

• Weibel instability:

Weibel ’59; … Mrowczynski ’88, ’93, ’94; Arnold, Lenaghan, Moore ‘03;

Romatschke, Strickland ‘03; very many since then…

oblate anisotropy ! Txx » Tyy À Tzz

• Nielsen-Olesen instability:

tiso » O(1/g1/2B1/2)

Nielsen, Olesen ’78; Chang, Weiss ’79; … Iwasaki ’08; Fujii, Itakura ’08 …

Short-time dynamics

• Anisotropy of the stress tensor Tij in a local rest frame:

“homogeneous“

background field

characteristic momentum of typical excitation

B z

z (force)

x (current)

y (magnetic field)

h{

,

}i

Nonperturbative description that contains all possible mechanisms:

Real-time lattice QCD

• Quantum simulation not possible so far (stochastic quantization?)

• Classical-statistical simulation: includes all fluctuations with

(includes, in particular, HTL / Vlasov)

physics of nonequilibrium instabilities ! high occupation numbers

• Classical simulations well tested for quantum evolutions in scalar theories:

p

Berges, Rothkopf, Schmidt,

PRL 101 (2008) 041603

(N=4)-component 4

parametric/spinodal

instability

i.e. “occupation numbers“ À 1h{ A(x) , A(y) }i À h[ A(x) , A(y) ]i

Classical-statistical lattice gauge field simulations

with À z (extreme anisotropy)

Normalized Gaussian probability functional

Here: = 0 / = s = 4, axial-temporal/Coulomb gauge

Wilson action:

Romatschke, Venugopalan (CGC); Berges, Gelfand, Scheffler, Sexty

Initial conditions:

P[A(t=0),tA(t=0)]: h A(t) A(t‘) i = s DA(0) DtA(0) P[A(0),tA(0)] A(t) A(t‘)

C is adjusted to obtain a given energy density , e.g. RHIC ~ 5-25 GeV/fm3

kT

/1/4 = 1

primary

secondary growth rates

Inverse primary growth rates:

Characteristic time scales

Berges, Scheffler, Sexty, PRD 77 (2008) 034504

e.g. RHIC ~ 5-25 GeV/fm3, LHC ~ 2 x RHIC

fast slow

fast: SU(2)

(Coulomb gauge)

Comparison SU(2) vs. SU(3)

) reduced primary growth rates by about 25% for given with /1/4 = 1

Berges, Gelfand, Scheffler, Sexty, PLB 677 (2009) 210

SU(3)

) Measured in units of the characteristic screening masses mT,SU(2) and

mT,SU(3), respectively, the primary growth rate is independent of Nc

Comparison SU(2) vs. SU(3)

) SU(3) / SU(2) = (3/4) £ ( SU(2) / SU(3) ) :

SU(2) = SU(3) i.e. 25% reduction, SU(2) / SU(3) = (8/3)1/4 i.e. only ~4%

C » 2

/1/4 = 1

• Initial conditions with faster isotropization/thermalization?

/ 1/4 ¿ 1 corresponds to rather “homogeneous“ field configurations (z / 1/4 ¿ 1)

pT

) growth rate increases

for smaller / 1/4

! Nielsen-Olesen?

SU(2)

Pressure

Fast bottom-up isotropization for momenta:

fast slow

pz . 1 GeV

Spatial Fourier transform of stress tensor T(t,x): PL(t,p) for ==3, PT(t,p)

/1/4 = 1

SU(2)

Kolmogorov wave turbulence

Berges, Scheffler, Sexty, arXiv:0811.4293 [hep-ph]

See however: Arnold, Moore PRD 73 (2006) 025006; Mueller, Shoshi, Wong, NPB 760 (2007) 145

• Different infrared behavior? Nonthermal IR fixed point?

(Infrared occupation number » 1/g2 ) strongly correlated)

• Scaling exponent close to the perturbative value = 4/3

t M

Kolmogorov

IR fixed point

Berges, Rothkopf, Schmidt, PRL 101 (2008) 041603

» k-2

» k-4 ?

Arnold, Moore `06

(Vlasov)

= 2 excluded!?

Berges, Scheffler, Sexty `08

(classical-statistical lattice QCD)

pe

rtu

rba

tive

(Ko

lmo

go

rov)

• Apparent discrepancy:

• Compare scalar instability dynamics (parametric resonance):

(perturbative)(nonperturbative)

not fully

understood

perturbatively

RG: ‘microscope‘ with varying resolution of length scale

~ 1/k

Fixed point: physics looks the same for ‘all‘ resolutions (in rescaled units)

scaling form, e.g. ‘length rescaled‘ correlator

Slow dynamics: Nonthermal RG fixed points

similarly, retarded propagator:

anomalous dimension

Typically not for all resolutions: • IR fixed point for k ! 0

• UV fixed point for k ! ‘1‘

‘occupation number‘ exponent

• vacuum:

• thermal:

• nonequilibrium:

Fluctuation-dissipation relation:

No fluctuation-dissipation relation:

spatial dimension d

dynamical exponent z

A) Hierarchy of infrared fixed point solutions for scalar N-component QFT:

Berges, Hoffmeister, NPB 813 (2009) 383

, ... (NLO 1/N)

Relativistic ( z ' 1) scalar inflaton in d = 3, ' 0 ) ' 4 , …

B) Ultraviolet fixed point solutions (classical-statistical limit, Kolmogorov):

Relativistic scalar inflaton in d = 3 ) ' 3/2 , 4/3 , …

No Kolmogorov turbulence in the far UV due to quantum corrections

Instability-induced fermion production

Quantum evolution of SU(2)L£SU(2)R linear sigma model (2PI 1/N to NLO):

• Fermion production proceeds with the

maximum primary boson growth rate!

Berges, Pruschke, Rothkopf, PRD 80 (2009) 023522

thermal equilibrium:

• Fast approach to Fermi-Dirac distribution

in the infrared!

Bosons

(spinodal)Fermions

V

g2 / = 1

fast slow

• Fast fermion growth induced via boson-fermion loop:

Boson

Fermion

(no IR scaling ! Pauli principle) • Scaling behavior at higher momenta

• Bosons practically unaffected for Yukawa couplings . O(1)

! Instability-induced fermion production can lead to substantial

deviations from standard production processes

1/s

» O(1)

Quantitatively: Classical-statistical lattice QCD with (quantum)

fermions can be simulated with well established techniques!

Conclusions

• Plasma instabilities for CGC type initial conditions ( / 1/4 = 1)

• ‘Bottom-up’ isotropization for p . 1 GeV,

i.e. (optimistically) about the range where hydro ‘works’

• (Perturbative) Kolmogorov wave turbulence with ' 4/3

Nonthermal IR fixed point in QCD?

• Initial conditions with faster isotropization/thermalization?

• Instability-induced fermion production can lead to substantial

deviations from standard production processes

! fast thermalization of low-momentum fermions on time scale

of maximum boson growth rate seen in linear sigma model

! » O(QCD) ?

-1 ' 1 – 2 fm/c